- name
- memory-network
- description
- Social memory network — discover communities and build trust through shared emotional memory, not follower counts or algorithms. Powered by Echo's memory graph infrastructure.
- metadata
- version
- 0.1.0
Memory Network
Social memory network. Discover communities and build trust through shared emotional memory — not follower counts, not algorithms, not curated personas.
How it works
Memory Network maps the social layer on top of individual memory graphs. When multiple people's memories resonate on the same experiences, emotions, or ideas, a network edge forms — creating organic communities grounded in genuine understanding.
Core capabilities
- Memory-based community discovery — clusters of people whose memory graphs share deep structural similarity, surfaced automatically
- Trust through transparency — every connection is backed by visible memory evidence; no black-box recommendations
- Cross-platform memory aggregation — pull memories from ChatGPT, Gemini, Claude, and other AI conversations into a unified social graph
- Real-time network evolution — as new memories form, network connections strengthen, weaken, or emerge organically
- Privacy-first sharing — users choose which memory clusters are public, which are private, and which are matchable
Architecture
Built on Echo's three-layer memory system:
- Identity layer — compressed profile (~500 tokens), always available
- Working memory — dynamic context per conversation (~5-10K tokens)
- Long-term storage — Supabase + HNSW vector search (unlimited)
Network edges are computed across Layer 3, with social context injected through Layer 2.
The vision
Every social platform today connects people through what they *show*. Memory Network connects people through what they *remember* — the experiences that shaped them, the ideas that moved them, the emotions they carry.
Part of the Echo ecosystem
Memory Network is a component of Echo Chat by Iditor — building memory as social identity infrastructure.
Status
Early development. Memory graph validated with 1,100+ memories, 23 clusters, emotion-aware matching active in beta (K-Factor 2, D7 retention 42%).